Overview

Use a common generation API with OpenAI, Anthropic, Google, and other supported providers. Model parameters, native tools, caching, and audio support vary by provider. See the 2.6.0 upgrade guide for new APIs and compatibility limits.

For the current beta, start with release availability and compatibility. Decisions, realtime voice and image generation pages describe the unreleased next-beta branch. The Feature Lab guide explains how to test each feature locally and what has been verified with real provider calls.

import { streamText } from "@yourgpt/llm-sdk";
import { openai } from "@yourgpt/llm-sdk/openai";

const result = await streamText({
  model: openai("gpt-4o"),
  prompt: "Hello!",
});

Switch Providers Instantly

Basic text generation shares the same call shape. Check feature-specific support before switching providers in workflows that use native tools or replay state.

import { openai } from '@yourgpt/llm-sdk/openai';

const model = openai('gpt-4o');

import { anthropic } from '@yourgpt/llm-sdk/anthropic';

const model = anthropic('claude-sonnet-4-20250514');

import { google } from '@yourgpt/llm-sdk/google';

const model = google('gemini-2.0-flash');

import { xai } from '@yourgpt/llm-sdk/xai';

const model = xai('grok-3-fast');

Framework Examples

Works with any HTTP framework. Returns standard Web Response objects.

app/api/chat/route.ts

import { streamText } from '@yourgpt/llm-sdk';
import { openai } from '@yourgpt/llm-sdk/openai';

export async function POST(req: Request) {
  const { messages } = await req.json();

  const result = await streamText({
    model: openai('gpt-4o'),
    messages,
  });

  return result.toTextStreamResponse();
}

server.ts

import express from 'express';
import { streamText } from '@yourgpt/llm-sdk';
import { openai } from '@yourgpt/llm-sdk/openai';

const app = express();
app.use(express.json());

app.post('/api/chat', async (req, res) => {
  const result = await streamText({
    model: openai('gpt-4o'),
    messages: req.body.messages,
  });

  const response = result.toTextStreamResponse();
  res.set(Object.fromEntries(response.headers));
  response.body?.pipeTo(new WritableStream({
    write(chunk) { res.write(chunk); },
    close() { res.end(); }
  }));
});

src/index.ts

import { Hono } from 'hono';
import { streamText } from '@yourgpt/llm-sdk';
import { openai } from '@yourgpt/llm-sdk/openai';

const app = new Hono();

app.post('/api/chat', async (c) => {
  const { messages } = await c.req.json();

  const result = await streamText({
    model: openai('gpt-4o'),
    messages,
  });

  return result.toTextStreamResponse();
});

export default app;

server.ts

import Fastify from 'fastify';
import { streamText } from '@yourgpt/llm-sdk';
import { openai } from '@yourgpt/llm-sdk/openai';

const app = Fastify();

app.post('/api/chat', async (req, reply) => {
  const result = await streamText({
    model: openai('gpt-4o'),
    messages: req.body.messages,
  });

  const response = result.toTextStreamResponse();
  reply.headers(Object.fromEntries(response.headers));
  return reply.send(response.body);
});

Core Functions

Function	Description
`streamText()`	Stream responses in real-time
`generateText()`	Generate complete responses
`tool()`	Define AI-callable functions
Parameters and fallback	Handle reasoning settings and specific compatibility failures
Responses	Preserve OpenAI conversation and tool state
Image tool results	Return native images to supported models
`generateSpeech()` / `transcribe()`	Generate speech and transcribe recorded audio
`experimental_decide()`	Native decisions; next beta, unreleased
`generateImage()`	Image generation and editing; next beta, unreleased
Realtime voice	Browser voice calls; next beta, experimental

Why Unified?

One API - Learn once, use everywhere
Swap providers - Change one import, keep all your code
Type-safe - Full TypeScript support across all providers
Standard responses - Web Response objects work anywhere